karylab_agents/vllm
1
0
Fork 0
forked from Karylab-cklius/vllm
Code Pull Requests 2 Actions 1 Packages Activity
17,137 Commits 354 Branches 175 Tags
cbfaaeceeb432f906ac074e3ede6cb63168eb897
Commit Graph
81 Commits
Author SHA1 Message Date
Yongye ZhuGitHubClaude Opus 4.8
 		cb7eb5c9b4 [MiniMax M3] Relocate sparse attention into the model dir + port decode top-k (#12) 
Move the MiniMax M3 sparse-attention backend and Triton kernels out of the
shared attention tree into the model definition (mirroring deepseek_v4), split
the kernels by role, and tighten the backend, then port the dedicated split-K
decode top-k from the sglang reference.

Structure:
- vllm/v1/attention/backends/minimax_m3_sparse.py
    -> vllm/models/minimax_m3/common/sparse_attention.py
- vllm/v1/attention/ops/minimax_m3_sparse_ops.py split into
    common/ops/index_topk.py  (index-score + top-k kernels)
    common/ops/sparse_attn.py (block-sparse GQA attention kernels)
  Pure-Triton/cross-platform, so under common/ (not nvidia/).
- registry enum + doc generator RELEVANT_PATTERNS repointed; the auto-generated
  attention_backends.md now has a dedicated "MiniMax M3 Sparse Attention" section.

Backend cleanups:
- Drop redundant metadata fields (sparse-selection params, decode max_query_len,
  decode cu_seqlens_q/context_lens) and read them from the layer/impl instead.
- Impl ctor takes explicit named args instead of kwargs.get lookups.
- Cudagraph support UNIFORM_SINGLE_TOKEN_DECODE -> UNIFORM_BATCH; reorder
  threshold and decode buffer scale by 1 + num_speculative_tokens.
- Unify the per-request context-length buffer (max_num_batched_tokens).
- Declare bf16-only KV cache.

Decode top-k (ported from sglang minimax_sparse_ops/decode):
- Split-K top-k: per-chunk partial top-k (_topk_index_partial_kernel) + merge
  (_topk_index_merge_kernel), replacing the single-program prefill top-k reuse.
- init/local block forcing moved into the decode score kernel (matches sglang).
- Verified against a torch reference (selected-block sets match across varied
  seq lengths, with/without init+local forcing, and short sequences).

Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-05-31 16:19:23 -04:00
Mohammad Miadh AngkadandGitHub 158289e0fc [Docs] Fix MLA prefill backend default docs (#43697) 
Signed-off-by: Mohammad Miadh Angkad <176301910+mmangkad@users.noreply.github.com>
 2026-05-27 10:13:22 +00:00
Bugen ZhaoGitHubNick HillEric CurtinDev-X25874Will.houWill.hou
 		39910f2b25 [Rust Frontend] Move code from vllm-frontend-rs (#43283) 
Signed-off-by: Bugen Zhao <i@bugenzhao.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Signed-off-by: Eric Curtin <eric.curtin@docker.com>
Signed-off-by: Dev-X25874 <283057883+Dev-X25874@users.noreply.github.com>
Signed-off-by: Will.hou <1205157517@qq.com>
Signed-off-by: Will.hou <willamhou@ceresman.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Eric Curtin <eric.curtin@docker.com>
Co-authored-by: Dev-X25874 <283057883+Dev-X25874@users.noreply.github.com>
Co-authored-by: Will.hou <1205157517@qq.com>
Co-authored-by: Will.hou <willamhou@ceresman.com>

Please see https://github.com/Inferact/vllm-frontend-rs for full original commit history.
 2026-05-21 17:21:48 -07:00
Lanze LiuGitHubMatthew Bonanni
 		b2c58ee942 [FlashAttn] Fix supports_kv_cache_dtype() accepting unhandled fp8 kv-cache dtype variants (#42685) 
Signed-off-by: Lanze Liu <lanzetech@gmail.com>
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-05-15 15:34:59 -04:00
Aaron HaoandGitHub e0a45f1455 [Feat][RL] IPC weight sync optimizations: multigpu support and chunked packed tensors (#37476) 
Signed-off-by: ahao-anyscale <ahao@anyscale.com>
Signed-off-by: hao-aaron <ahao@anyscale.com>
 2026-05-15 22:53:06 +08:00
CynicDoraandGitHub 256dbcaabf [Feature] Support custom callable proposer backend for speculative decoding (#39487) 
Signed-off-by: 524031910363 <hyzhyzsh@sjtu.edu.cn>
Signed-off-by: CynicDora <hyzhyzsh@sjtu.edu.cn>
 2026-05-13 16:53:01 +00:00
Matthew BonanniandGitHub be5983b874 [Docs] Add non-causal support to attention backend docs (#41643) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-05-04 20:35:15 +00:00
Matthew BonanniGitHubMichael GoinLucas Wilkinsonmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 		f3fef12350 [Attention] Abstract the MLA prefill backends and eliminate cuDNN (#32623) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-05-01 13:36:20 -04:00
sychen52andGitHub 947138b6c2 Add nvfp4 kv cache support (#40177) 
Signed-off-by: Shiyang Chen <shiychen@nvidia.com>
 2026-05-01 04:55:16 +00:00
Jiangyun ZhuandGitHub e8ee2a78db [Attention] use diff kv backend for mimo v2 flash (#40045) 
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
 2026-04-24 11:25:55 +00:00
Martin HickeyandGitHub 3951d3eacd [MyPy] Enable mypy for vllm/model_executor/layers/ (#40159) 
Signed-off-by: Martin Hickey <martin.hickey@ie.ibm.com>
 2026-04-21 20:15:02 -07:00
Vadim GimpelsonandGitHub 6d85b36a9f Revert #38730 and #38791 (#40032) 
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
Signed-off-by: Vadim Gimpelson <156319763+vadiklyutiy@users.noreply.github.com>
 2026-04-21 11:44:11 -04:00
Harry MellorGitHubRobert Shaw
 		fc645f1acc Add structure to requirements/ directory (#39024) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
 2026-04-10 13:46:41 -07:00
Yan MaandGitHub ec68d53b2b Add platform manual_seed_all API (#38468) 
Signed-off-by: Yan Ma <yan.ma@intel.com>
 2026-04-10 13:43:50 +08:00
Wentao YeandGitHub aec18492d0 [CI] Fix mypy for vllm/v1/ops (#39219) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2026-04-09 11:06:34 +08:00
Stefano CastagnettaandGitHub 6183cae1bd [Bugfix] Restrict TRTLLM attention to SM100, fixing GB300 (SM103) hang (#38730) 
Signed-off-by: Stefano Castagnetta <scastagnetta@nvidia.com>
 2026-04-01 12:08:40 -07:00
wliao2GitHubHarry MellorKunshang Ji
 		4dfad17ed1 replace cuda_device_count_stateless() to current_platform.device_count() (#37841) 
Signed-off-by: Liao, Wei <wei.liao@intel.com>
Signed-off-by: wliao2 <wei.liao@intel.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
 2026-03-31 22:32:54 +08:00
Lucas KabelaandGitHub e31915063d [Bugfix] Fix for builtins (forward fix of pytorch/177558) (#37234) 
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
 2026-03-31 01:08:11 +00:00
Andreas KaratzasandGitHub 43cc5138e5 [ROCm][CI] Fix cross-attention dispatch for encoder-decoder models (#38450) 
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
 2026-03-28 22:08:03 -07:00
Harry MellorandGitHub b3601da6e7 [Mypy] Fix mypy for vllm/model_executor (except vllm/model_executor/layers) (#37904) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-03-24 17:14:01 +00:00
Wentao YeandGitHub 45bd5c8e75 [Mypy] Fix mypy for vllm/config (#37808) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2026-03-23 14:33:59 +00:00
Wei ZhaoandGitHub b36adfa349 [Perf] Set Flashinfer sparse MLA as default backend for FP8 kv cache (#37252) 
Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
 2026-03-17 20:09:20 +00:00
Isotr0pyandGitHub a836524d20 [Chore] Replace all base64 usages with faster pybase64 package (#37290) 
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
 2026-03-17 14:44:19 +00:00
Kunshang JiandGitHub 747b068136 [Hardware] Replace memory related torch.cuda APIs (#37031) 
Signed-off-by: Kunshang Ji <jikunshang95@gmail.com>
 2026-03-16 10:24:48 +00:00
Dimitrios BariamisGitHubDimitrios Bariamis
 		cc16b24b17 Update Flashinfer to 0.6.6 (#36768) 
Signed-off-by: Dimitrios Bariamis <12195802+dbari@users.noreply.github.com>
Co-authored-by: Dimitrios Bariamis <12195802+dbari@users.noreply.github.com>
 2026-03-12 13:19:19 -04:00
Kunshang JiandGitHub 53ec16a705 [Hardware] Replace torch.cuda.device_count/current_device/set_device API (#36145) 
Signed-off-by: Kunshang Ji <jikunshang95@gmail.com>
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
 2026-03-12 07:57:47 -07:00
Martin HickeyGitHubHarry Mellor
 		7f1f36bf91 [CI] Fix mypy for vllm/reasoning (#35742) 
Signed-off-by: Martin Hickey <martin.hickey@ie.ibm.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-03-12 12:21:33 +00:00
Yan MaandGitHub 894843eb25 replace with torch.cuda.device with with torch.accelerator.device_index (#36144) 
Signed-off-by: Yan Ma <yan.ma@intel.com>
 2026-03-11 23:12:57 -07:00
Harry MellorandGitHub a0f44bb616 Allow markdownlint to run locally (#36398) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-03-08 20:05:24 -07:00
Wei ZhaoandGitHub 379689d533 [Perf] Support FP8 KV cache for Flashinfer MLA Sparse (#35891) 2026-03-07 13:51:54 -08:00
Kunshang JiandGitHub 66a2209645 [Hardware] Replace torch.cuda.synchronize() api with torch.accelerator.synchronize (#36085) 
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
 2026-03-05 10:36:39 +00:00
Taneem IbrahimGitHubHarry Mellor
 		1aaec59d79 [MISC] fixed tool_parser mypy errors (#35640) 
Signed-off-by: Taneem Ibrahim <taneem.ibrahim@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-03-04 12:23:12 +00:00
Kunshang JiGitHubHarry Mellor
 		16d2ad1d38 [Hardware] Replace torch.cuda.empty_cache with torch.accelerator.empty_cache (#30681) 
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
Signed-off-by: Kunshang Ji <jikunshang95@gmail.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-03-04 09:49:47 +00:00
Martin HickeyGitHubHarry Mellor
 		7560d674c9 [CI] Fix mypy for vllm/device allocator (#35518) 
Signed-off-by: Martin Hickey <martin.hickey@ie.ibm.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-03-02 15:53:18 +00:00
Lucas WilkinsonGitHubMatthew BonanniTyler Michael Smith
 		8b5014d3dd [Attention] FA4 integration (#32974) 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
 2026-03-01 23:44:57 +00:00
Taneem IbrahimandGitHub 59d7af9c6c [MISC] Fixing a null reference by removing parallel_utils from mypy EXCLUDE (#35630) 
Signed-off-by: Taneem Ibrahim <taneem.ibrahim@gmail.com>
 2026-03-01 09:26:44 -05:00
Aaron HaoandGitHub 2ce6f3cf67 [Feat][RL][2/2] Native Weight Syncing API: IPC (#34171) 
Signed-off-by: hao-aaron <ahao@anyscale.com>
Signed-off-by: Aaron Hao <ahao@anyscale.com>
Signed-off-by: ahao-anyscale <ahao@anyscale.com>
 2026-02-27 13:45:21 -07:00
Taneem IbrahimandGitHub d38cd3dde5 [Misc] Fix mypy errors in vllm/profiler and remove from exclude list (#34959) 
Signed-off-by: Taneem Ibrahim <taneem.ibrahim@gmail.com>
 2026-02-20 19:56:33 -08:00
junuxyzandGitHub c61a98f529 [CI][BugFix] ShellCheck cleanup to remove baseline and preserve runtime behavior (#34514) 
Signed-off-by: junuxyz <216036880+junuxyz@users.noreply.github.com>
 2026-02-17 12:22:56 +00:00
Aneesh PutturandGitHub 0b5f9b7204 [CI] Enable mypy import following for vllm/v1/kv_offload (#34639) 
Signed-off-by: Aneesh Puttur <aneeshputtur@gmail.com>
 2026-02-17 09:58:15 +08:00
Lucas KabelaGitHubHarry Mellor
 		a3205beffb [CI] Enable mypy coverage for individual excluded files (#34292) 
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-02-16 07:34:29 -08:00
Matthew BonanniGitHubLucas WilkinsonLucas Wilkinson
 		f2c47886fd [Attention] Add FlashInfer Sparse MLA backend (#33451) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
 2026-02-12 17:21:54 +00:00
junuxyzandGitHub fa7e0bfacf [CI][BugFix] Fix silent failure in shellcheck hook and baseline exist… (#32458) 
Signed-off-by: junuxyz <216036880+junuxyz@users.noreply.github.com>
 2026-02-11 17:03:48 +00:00
Tyler Michael SmithGitHubClaude Opus 4.6
 		c4b9e6778f [Misc] Add pre-commit hook to catch boolean ops in with-statements (#34271) 
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
 2026-02-10 15:13:20 -08:00
Michael GoinandGitHub 5e75a14a66 [Doc] Add DCP support to attention backend doc (#33936) 2026-02-09 18:33:43 -05:00
Harry MellorandGitHub 791a94bed0 Consolidate and fix forbidden import pre-commit checks (#33982) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-02-06 01:47:41 -08:00
Harry MellorandGitHub 61e632aea1 Turn @config into a dataclass_transform (#31541) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-02-03 17:40:59 +00:00
Lucas KabelaandGitHub 726d89720c [CI] Enable mypy import following for vllm/spec_decode (#33282) 
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
 2026-01-30 06:43:32 +00:00
Harry MellorandGitHub fb946a7f89 Make mypy opt-out instead of opt-in (#33205) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2026-01-29 09:12:26 +00:00
Matthew BonanniandGitHub 77c4f45c6c [7/N][Attention][Docs] Add documentation for attention backends (#32477) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-01-28 17:20:22 -05:00